If your review "of a book" is actually a review of the author, I agree that it's not quite right to call it a fake review, nor fraud, but neither is it truly a review of the book. Ideally there would be a way to rate authors so that people don't feel compelled to categorize their author reviews as book reviews for books they haven't read. People are going to leave these kind of reviews, so the system may as well be set up for it.
Generally, I'm wary of attempts to address 'review bombing' (besides standard anti-botting measures) because it seems like accusations of review bombing have become a sort of general cope employed by creators whenever their thing gets a negative reception. It's common to de-legitimize contrary points of view by calling people dehumanizing terms like 'troll', when in fact those people probably earnestly feel that way (they aren't trolling.) If I know the earth is round but I pretend to think the earth is flat to get a rise out of people, that's trolling. But if I'm a halfwit who earnestly thinks the earth is flat, and I say so, that's not trolling. The difference between these won't necessarily be apparent from the text of the review itself; how then do you separate the trolls from the people you disagree with? On an individual basis you can go with your gut or look that person up, but that doesn't scale up to classifying thousands of reviews. What creators are really asking for is the privilege to curate the reviews themselves, but that isn't a reasonable privilege to grant because it would completely devalue reviews for consumers.
Do what needs to be done to combat botting, use captchas or account verification or whatever works. But after that? Let the reviews fall where they may. Creators will cry that the bad reviews aren't legitimate because the reviewers are [dehumanizing term], but there's nothing that can reasonably be done about this.